Medical physics最新文献

{"title":"Boosting 2D brain image registration via priors from large model.","authors":"Hao Lin, Yonghong Song","doi":"10.1002/mp.17696","DOIUrl":"https://doi.org/10.1002/mp.17696","url":null,"abstract":"<p><strong>Background: </strong>Deformable medical image registration aims to align image pairs with local differences, improving the accuracy of medical analyses and assisting various diagnostic scenarios.</p><p><strong>Purpose: </strong>We aim to overcome these challenges: Deep learning-based registration approaches have greatly enhanced registration speed and accuracy by continuously improving registration networks and processes. However, the lack of extensive medical datasets limits the complexity of registration models. Optimizing registration networks within a fixed dataset often leads to overfitting, hindering further accuracy improvements and reducing generalization capabilities.</p><p><strong>Methods: </strong>We explore the application of the foundational model DINOv2 to registration tasks, leveraging its prior knowledge to support learning-based unsupervised registration networks and overcome network bottlenecks to improve accuracy. We investigate three modes of DINOv2-assisted registration, including direct registration architecture, enhanced architecture, and refined architecture. Additionally, we study the applicability of three feature aggregation methods-convolutional interaction, direct fusion, and cross-attention-within the proposed DINOv2-based registration frameworks.</p><p><strong>Results: </strong>We conducted extensive experiments on the IXI and OASIS public datasets, demonstrating that the enhanced and refined architectures notably improve registration accuracy, reduce data dependency, and maintain strong generalization capabilities.</p><p><strong>Conclusion: </strong>This study offers novel approaches for applying foundational models to deformable image registration tasks.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Self-supervised arbitrary-scale super-angular resolution diffusion MRI reconstruction.","authors":"Shuangxing Wang, Lihui Wang, Ying Cao, Zeyu Deng, Chen Ye, Rongpin Wang, Yuemin Zhu, Hongjiang Wei","doi":"10.1002/mp.17691","DOIUrl":"https://doi.org/10.1002/mp.17691","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Diffusion magnetic resonance imaging (dMRI) is currently the unique noninvasive imaging technique to investigate the microstructure of in vivo tissues. To fully explore the complex tissue microstructure at sub-voxel scale, diffusion weighted (DW) images along many diffusion gradient directions are usually acquired, this is undoubtedly time consuming and inhibits their clinical applications. How to estimate the tissue microstructure only from DW images acquired with few diffusion directions remains a challenge.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;To address this challenge, we propose a self-supervised arbitrary scale super-angular resolution diffusion MRI reconstruction network (SARDI-nn), which can generate DW images along any directions from few acquisitions, allowing to overcome the limits of diffusion direction number on exploring the tissue microstructure.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;SARDI-nn is mainly composed of a diffusion direction-specific DW image feature extraction (DWFE) module and a physics-driven implicit expression and reconstruction (IRR) module. During training, dual downsampling operations are implemented. The first downsampling is used to produce the low-angular resolution (LAR) DW images; the second downsampling is for constructing input and learning target of SARDI-nn. The input LAR DW images pass through a DWFE module (composed of several residual blocks) to extract the feature representations of DW images along input directions, and then these features and the difference between the any querying diffusion direction and the input directions are input into a IRR module to derive the implicit representation and DW image along this query direction. Finally, based on the principle of dMRI, an adaptive weighting method is used to refine the DW image quality. During testing, given any diffusion directions, we can simply infer the corresponding DW images along these directions, accordingly, SARDI-nn can realize arbitrary scale angular super resolution. To test the effectiveness of the proposed method, we compare it with several existing methods in terms of peak signal to noise ratio (PSNR), structural similarity index measure (SSIM), and root mean square error (RMSE) of DW image and microstructure metrics derived from diffusion kurtosis imaging (DKI) and neurite orientation dispersion and density imaging (NODDI) models at different upsampling scales on Human Connectome Project (HCP) and several in-house datasets.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The comparison results demonstrate that our method achieves almost the best performance at all scales, with SSIM of reconstructed DW images improved by 10.04% at the upscale of 3 and 5.9% at the upscale of 15. Regarding the microstructures derived from DKI and NODDI models, when the upscale is not larger than 6, our method outperforms the best supervised learning method. In addition, the test results on external datasets show the well generality of our method.","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143460005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Secondary cancer risk in head-and-neck cancer patients: A comparison of RBE-weighted proton therapy and photon therapy.","authors":"Peter Dasiukevich, Sebastian Tattenberg, Cornelia Hoehr, Abdelkhalek Hammi","doi":"10.1002/mp.17705","DOIUrl":"https://doi.org/10.1002/mp.17705","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Secondary cancer is a serious side effect from external beam radiotherapy (EBRT). Conventional EBRT is performed using a beam of photons, however, due to their ability to produce more conformal dose distributions, the use of protons is becoming more wide-spread. Due to this sparing it would be expected that proton therapy could be associated with lower secondary cancer rates compared to photon therapy. However, since proton therapy data is still being accumulated and the follow-up period is often relatively short thus far, simulation studies can complement the existing data and extrapolate to longer time frames.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;This study aims to estimate and compare the risk of secondary cancer when treating head-and-neck cancer patients with proton therapy or photon therapy, while combining a whole-body computational human phantom with the patient treatment planning computed tomography (CT) scan in order to study organs that are partially or fully outside of the treatment planning CT. In addition, proton therapy secondary cancer rates are investigated further by including variable relative biological effectiveness (RBE) models.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;For 20 head-and-neck cancer patients, two clinical radiotherapy treatment plans were created, one for proton therapy and one for photon therapy. For proton therapy, linear energy transfer (LET) distributions were simulated and used to calculate the variable RBE-weighted dose distributions for six different variable RBE models, in addition to the constant RBE of 1.1 widely used clinically. In order to obtain the dose deposited outside the treatment planning CT scan, an adjustable whole-body digital reference phantom was stitched to the treatment planning CT. Based on the resulting dose distributions, the risk of secondary cancer was calculated for each modality.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Averaged across all patients and relevant organs, photon therapy compared to proton therapy with a constant RBE of 1.1 was estimated to be 1.8 times more likely to cause secondary cancer. This risk ratio varied between 1.6 and 2.0, depending on the variable RBE model used. Cases with lifetime attributable risk (LAR) values below 0.1% were excluded from this analysis to prevent the benefits of proton therapy (the ratio &lt;math&gt; &lt;semantics&gt; &lt;mfrac&gt;&lt;mrow&gt;&lt;mi&gt;L&lt;/mi&gt; &lt;mi&gt;A&lt;/mi&gt; &lt;msub&gt;&lt;mi&gt;R&lt;/mi&gt; &lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mi&gt;h&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;t&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;n&lt;/mi&gt;&lt;/mrow&gt; &lt;/msub&gt; &lt;/mrow&gt; &lt;mrow&gt;&lt;mi&gt;L&lt;/mi&gt; &lt;mi&gt;A&lt;/mi&gt; &lt;msub&gt;&lt;mi&gt;R&lt;/mi&gt; &lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mi&gt;r&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;t&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;n&lt;/mi&gt;&lt;/mrow&gt; &lt;/msub&gt; &lt;/mrow&gt; &lt;/mfrac&gt; &lt;annotation&gt;$frac{LAR_{photon}}{LAR_{proton}}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; ) from being artificially elevated in cases in which &lt;math&gt; &lt;semantics&gt;&lt;mrow&gt;&lt;mi&gt;L&lt;/mi&gt; &lt;mi&gt;A&lt;/mi&gt; &lt;msub&gt;&lt;mi&gt;R&lt;/mi&gt; &lt;mrow&gt;&lt;mi&gt;p&lt;/mi&gt; &lt;mi&gt;r&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;t&lt;/mi&gt; &lt;mi&gt;o&lt;/mi&gt; &lt;mi&gt;n&lt;/mi&gt;&lt;/mrow&gt; &lt;/msub&gt; &lt;mo&gt;≈&lt;/mo&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/mrow&gt; &lt;annotation&gt;$LAR_{pro","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of secondary radiation dose between pencil beam scanning and scattered delivery for proton and VHEE radiotherapy.","authors":"Maria Grazia Ronga, Flavia Gesualdi, Anthony Bonfrate, Annalisa Patriarca, Régis Ferrand, Gilles Créhange, Irène Buvat, Ludovic De Marzi","doi":"10.1002/mp.17700","DOIUrl":"https://doi.org/10.1002/mp.17700","url":null,"abstract":"<p><strong>Background: </strong>Very high-energy electrons (VHEEs) in radiotherapy may offer several potential advantages over conventional electron beams and other techniques, for example, the fact that they can be used at ultra-high dose rates (UHDRs), therefore enabling FLASH radiotherapy. However, the production of secondary particles at high energies (50-200 MeV) has yet to be studied in detail for this technique currently under development.</p><p><strong>Purpose: </strong>The aim of this work was to examine the secondary dose produced by VHEEs, with particular emphasis on bremsstrahlung photons and neutrons, for two beam delivery systems (double scattering [DS] and pencil beam scanning [PBS]).</p><p><strong>Methods: </strong>The electron, X-ray, and neutron doses arising from two beam delivery systems (DS or PBS) were computed using Monte Carlo (MC) simulations in the TOPAS (TOol for PArticle Simulation)/Geant4 toolkit, and a preliminary assessment of the secondary dose for a clinical VHEE treatment was performed using a whole-body phantom. An evaluation of the secondary dose produced by this preliminary design of a VHEE nozzle set in a clinical proton facility was performed, taking into account realistic PBS or DS nozzle configurations.</p><p><strong>Results: </strong>The mean doses received by a patient undergoing DS-VHEE irradiation were found to be up to 5.3-fold and 6.8-fold higher for in-field or out-of-field organs for photons and neutrons, respectively, compared to the PBS-VHEE plan. The results for the secondary neutron dose in intracranial treatments also demonstrate the characteristic of VHEE compared to proton beams for reducing the out-of-field secondary neutron dose. The dose to the public area that could be delivered to meet regulatory limits surrounding a possible treatment room in a proton therapy facility was assessed. A regulatory limit of 0.5 µSv/h would give a restriction of 49 and 83 Gy per patient and per fraction for DS and PBS, respectively.</p><p><strong>Conclusions: </strong>This work describes a method to simulate and compare secondary radiation doses resulting from scattered, scanned VHEE or proton therapy treatments. The results indicate that a conventionally shielded proton therapy room results in acceptable public doses for a preliminary VHEE design and could be of interest for radiation protection purposes and for similar setups. Other facilities with differing layouts may, however, lead to different conclusions, requiring further studies.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461307","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantifying photon counting detector (PCD) performance using PCD-CT images.","authors":"Linying Zhan, Guang-Hong Chen, Ke Li","doi":"10.1002/mp.17701","DOIUrl":"https://doi.org/10.1002/mp.17701","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Photon counting detector CTs (PCD-CTs) have recently been introduced to clinical imaging. This development creates a new need for end-users to quantify and monitor the physical performance of PCDs. Traditionally, the characterization of PCD performance relied on detector counts, which are typically accessible to the manufacturer but are not usually available to clinical end-users.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;The goal of this work was to develop a new method for quantifying PCD performance using reconstructed PCD-CT images, without requiring access to the PCD counts.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;The proposed method is based on a set of closed-form relationships that connect PCD-CT image noise, the PCD deadtime ( &lt;math&gt;&lt;semantics&gt;&lt;mi&gt;τ&lt;/mi&gt; &lt;annotation&gt;$tau$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; ), and the zero-frequency detective quantum efficiency ( &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; ) of PCDs. At a low tube current (mA) level, the mean output counts of the PCD were estimated by fitting the measured PCD-CT noise power spectrum (NPS) to a parametric model. &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; was then calculated by normalizing the estimated mean detector counts to the expected input x-ray photon number. To estimate &lt;math&gt;&lt;semantics&gt;&lt;mi&gt;τ&lt;/mi&gt; &lt;annotation&gt;$tau$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; , the image variance of PCD-CT was measured at different mA levels. A novel quantitative relationship between PCD-CT image variance, &lt;math&gt;&lt;semantics&gt;&lt;mi&gt;τ&lt;/mi&gt; &lt;annotation&gt;$tau$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; , and mA was employed to estimate &lt;math&gt;&lt;semantics&gt;&lt;mi&gt;τ&lt;/mi&gt; &lt;annotation&gt;$tau$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; through parametric fitting. The method was validated using both simulated and experimental PCD-CT data, covering a range of &lt;math&gt;&lt;semantics&gt;&lt;mi&gt;τ&lt;/mi&gt; &lt;annotation&gt;$tau$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; , &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; , and system geometries.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;For the simulated curved-detector PCD-CT, the estimation errors for &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; and deadtime were -3.7% and 0.5%, respectively. For the simulated collinear-detector PCD-CT, the estimation errors for &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; and deadtime were -3.3% and -1.0%, respectively. For the experimental collinear-detector PCD-CT, the estimation errors for &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;DQE&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;${rm DQE}_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; and deadtime were -2.6% and 1.6%, respectively.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;By analyzing the variance and NPS of PCD-CT images, &lt;math&gt; &lt;semant","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143461309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Advanced prediction of multi-leaf collimator leaf position using artificial neural network.","authors":"Jun Lv, Liuli Chen, Zhiqiang Zhu, Pengcheng Long, Liqin Hu, Han Zhou, Zetian Shen","doi":"10.1002/mp.17690","DOIUrl":"https://doi.org/10.1002/mp.17690","url":null,"abstract":"<p><strong>Background: </strong>Multi-leaf collimators (MLCs) are crucial for modern radiotherapy as they ensure precise target irradiation through accurate leaf positioning. Accurate prediction of MLC leaf positions is vital for the effectiveness and safety of treatments.</p><p><strong>Purpose: </strong>This study aims to establish three neural network models for predicting the delivered positions of MLCs in radiotherapy.</p><p><strong>Methods: </strong>Fifty plans with sliding window dynamic intensity-modulated radiation therapy delivery were selected from an Elekta linear accelerator, which features a 160-leaf MLC system. The dose fraction, gantry angle, collimator angle, X1 and X2 jaw positions, Y1 and Y2 carriage positions, planned leaf positions, adjacent leaf positions, leaf gap, leaf velocity, and leaf acceleration were extracted from the planning data in the machine's log files and used as model inputs, with the delivered leaf positional serving as the target response. This establishes the input-output relationship for the neural network, and the predicted MLC positions are obtained through training. Particle Swarm Optimization Back Propagation Neural Network (PSOBPNN), Back Propagation Neural Network (BPNN), and Radial Basis Function Neural Network (RBFNN) architectures were developed to predict MLC leaf positional deviations during treatment. The training was conducted on 70% of the sample data, with the remaining 30% used for validation and testing. Model performance was assessed using metrics such as mean absolute error (MAE), mean squared error (MSE), regression plots, and error histograms.</p><p><strong>Results: </strong>The proposed neural network models demonstrated high accuracy in predicting MLC leaf positions. The PSOBPNN model demonstrated superior performance with an MAE of 0.0043 mm and an MSE of 0.00003 mm². In comparison, the BPNN model achieved an MAE of 0.0241 mm and an MSE of 0.001 mm², while the RBFNN model exhibited an MAE of 0.0331 mm and an MSE of 0.0019 mm². The correlation coefficient (R = 0.9999) of models indicates a close match between predicted and delivered leaf positions for all MLC leaves.</p><p><strong>Conclusion: </strong>Three models were evaluated for predicting the delivered MLC positions using data from an Elekta accelerator. The PSOBPNN model exhibited superior performance by achieving markedly lower MAE and MSE values while also demonstrating robust generalizability in predicting positions across various leaf indices, outperforming the conventional BPNN and RBFNN models.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143451269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Biological dose-based fractional dose optimization of Bragg peak FLASH-RT for lung cancer treatment.","authors":"Yiling Zeng, Qi Zhang, Wei Wang, Xu Liu, Bin Qin, Bo Pang, Muyu Liu, Shuoyan Chen, Hong Quan, Yu Chang, Zhiyong Yang","doi":"10.1002/mp.17697","DOIUrl":"https://doi.org/10.1002/mp.17697","url":null,"abstract":"<p><strong>Background: </strong>The FLASH effect is dose-dependent, and fractional dose optimization may enhance it, improving normal tissue sparing.</p><p><strong>Purpose: </strong>This study investigates the performance of fractional dose optimization in enhancing normal tissue sparing for Bragg peak FLASH radiotherapy (FLASH-RT).</p><p><strong>Methods: </strong>15 lung cancer patients, including eight with peripherally located tumors and seven with centrally located tumors, were retrospectively analyzed. A uniform fractionation prescription of 50 Gy in five fractions was utilized, corresponding to a biological equivalent dose (BED) of 100 Gy, calculated using an α/β value of 10 Gy. For each patient, uniform (UFD) and nonuniform fractional dose (non-UFD) plans were designed. In UFD FLASH plans, five multi-energy Bragg peak beams were optimized using single-field optimization, each delivering 10 Gy to the target. In non-UFD FLASH plans, fractional doses were optimized to enhance sparing effects while ensuring the target received a BED comparable to UFD plans. A dose-dependent FLASH enhancement ratio (FER) was integrated with the BED to form the FER-BED metric to compare the UFD and non-UFD plans. An α/β value of 3 Gy was applied for normal tissues in the calculations.</p><p><strong>Results: </strong>Bragg peak FLASH plans showed high dose conformality for both peripheral and central tumors, with all plans achieving a conformality index (the ratio of the volume receiving the prescribed dose to the CTV volume) below 1.2. In non-UFD plans, fractional doses ranged from 5.0 Gy to 20.0 Gy. Compared to UFD plans, non-UFD plans achieved similar BED coverage (BED_98%: 96.6 Gy vs. 97.1 Gy, p = 0.256), while offering improved organ-at-risk sparing. Specifically, the FER-BED_15cc for the heart reduced by 10.5% (9.4 Gy vs. 10.5 Gy, p = 0.017) and the V_6.7GyFER-BED for the ipsilateral lung decreased by 4.3% (29 .1% vs. 30.4%, p = 0.008). No significant difference was observed in FER-BED_0.25cc of spinal cord (UFD: 7.1 Gy, non-UFD: 6.9 Gy, p = 0.626) and FER-BED_5cc in esophagus (UFD: 0.4 Gy, non-UFD: 0.4 Gy, p = 0.831).</p><p><strong>Conclusions: </strong>Bragg peak FLASH-RT achieved high dose conformality for both peripheral and central tumors. Fractional dose optimization, using a single beam per fraction delivery mode, enhanced normal tissue sparing by leveraging both fractionation and FLASH effects.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143451270","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dose prediction via deep learning to enhance treatment planning of lung radiotherapy including simultaneous integrated boost techniques.","authors":"Wenhua Cao, Mary Gronberg, Stephen Bilton, Hana Baroudi, Skylar Gay, Christopher Peeler, Zhongxing Liao, Thomas J Whitaker, Karen Hoffman, Laurence E Court","doi":"10.1002/mp.17692","DOIUrl":"https://doi.org/10.1002/mp.17692","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Recent studies have shown deep learning techniques are able to predict three-dimensional (3D) dose distributions of radiotherapy treatment plans. However, their use in dose prediction for treatments with varied prescription doses including simultaneous integrated boost (SIB), that is, using multiple prescription doses within the same plan, and benefit in improving plan quality should be validated.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;To investigate the feasibility and potential benefit of using deep learning to predict dose distribution of volumetric modulated arc therapy (VMAT) including SIB techniques and improve treatment planning for patients with lung cancer.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;The dose prediction model was trained with 93 retrospective clinical VMAT plans for patients with lung cancer from our institutional patient database. The prescription doses of these plans ranged from 35 to 72 Gy, with various fractionation schemes. We used a 3D U-Net architecture to predict 3D dose distributions with 75 plans for training and 18 plans for testing. Model input consisted of computed tomography (CT) images, target and normal tissue contours and prescription doses. We first evaluated model accuracy by comparing the predicted and clinical plan doses for the test set, and then performed replanning according to predicted dose distributions. Furthermore, we evaluated the model prospectively in an additional set of 10 patients from our institution by two approaches where dose prediction was either blinded or provided to treatment planners. We then assessed whether dose prediction could identify suboptimal plan quality and how it affects plan quality if adopted in clinical planning workflow.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The dose prediction model achieved good agreement between the predicted and clinical plan dose distributions, with a mean dose difference of -0.49 ± 0.54 Gy across the test set. The replanning study guided by dose prediction showed that a small subset of the original plans could benefit from improvements regarding sparing of the spinal cord and esophagus. The analysis of the prospective dataset, with initial and final clinical plans generated in the absence of dose prediction, showed that the predicted doses were able to identify possible improvements of target coverage and normal tissue sparing in the initial plans similar to those made by the final plans for majority of the patients, but in varied magnitudes. Moreover, the plans generated with dose prediction guidance were able to consistently improve normal tissue sparing compared to the plans generated without dose prediction guidance.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;We demonstrated that our deep learning model can consistently predict high quality VMAT lung plans for a variety of prescription doses. The dose prediction tool was also effective in identifying suboptimal plan quality, suggesting its potential benefit in automated treatment planni","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143451271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing automated right-sided early-stage breast cancer treatments via deep learning model adaptation without additional training.","authors":"Michele Zeverino, Silvia Fabiano, Wendy Jeanneret-Sozzi, Jean Bourhis, Francois Bochud, Raphaël Moeckli","doi":"10.1002/mp.17682","DOIUrl":"https://doi.org/10.1002/mp.17682","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Input data curation and model training are essential, but time-consuming steps in building a deep-learning (DL) auto-planning model, ensuring high-quality data and optimized performance. Ideally, one would prefer a DL model that exhibits the same high-quality performance as a trained model without the necessity of undergoing such time-consuming processes. That goal can be achieved by providing models that have been trained on a given dataset and are capable of being fine-tuned for other ones, requiring no additional training.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;To streamline the process for producing an automated right-sided breast (RSB) treatment planning technique adapting a DL model originally trained on left-sided breast (LSB) patients via treatment planning system (TPS) specific tools only, thereby eliminating the need for additional training.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;The adaptation process involved the production of a predicted dose (PD) for the RSB by swapping from left-to-right the symmetric structures in association with the tuning of the initial LSB model settings for each of the two steps that follow the dose prediction: the predict settings for the post-processing of the PD (ppPD) and the mimic settings for the dose mimicking, respectively. Thirty patients were involved in the adaptation process: Ten manual plans were chosen as ground truth for tuning the LSB model settings, and the adapted RSB model was validated against 20 manual plans. During model tuning, PD, ppPD, and mimicked dose (MD) were iteratively compared to the manual dose according to the new RSB model settings configurations. For RSB model validation, only MD was involved in the planning comparison. Subsequently, the model was applied to 10 clinical patients. Manual and automated plans were compared using a site-specific list of dose-volume requirements.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;PD for the RSB model required substantial corrections as it differed significantly from manual doses in terms of mean dose to the heart (+11.1 Gy) and right lung (+4.4 Gy), and maximum dose to the left lung (+6.4 Gy) and right coronary (+11.5 Gy). Such discrepancies were first addressed by producing a ppPD always superior to the manual dose by changing or introducing new predict settings. Second, the mimic settings were also reformulated to ensure a MD not inferior to the manual dose. The final adapted version of the RSB model settings, for which MD was found to be not significantly different than the manual dose except for a better right lung sparing (-1.1 Gy average dose), was retained for the model validation. In RSB model validation, a few significant-yet not clinically relevant-differences were noted, with the right lung being more spared in auto-plans (-0.6 Gy average dose) and the maximum dose to the left lung being lower in the manual plans (-0.8 Gy). The clinical plans returned dose distributions not significantly different than the validation p","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143451286","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Measuring temperature in polyvinylpyrrolidone (PVP) solutions using MR spectroscopy.","authors":"Neville D Gai, Ruifeng Dong, Jan Willem van der Veen, Ronald Ouwerkerk, Carlo Pierpaoli","doi":"10.1002/mp.17683","DOIUrl":"https://doi.org/10.1002/mp.17683","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Polyvinylpyrrolidone (PVP) water solutions could be used for cross-site and cross-vendor validation of diffusion-related measurements. However, since water diffusivity varies as a function of temperature, knowing the temperature of the PVP solution at the time of the measurement is fundamental in accomplishing this task.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;MR spectroscopy (MRS) could provide absolute temperature measurements since the water peak moves relative to any stable peak as temperature changes. In this work, the PVP proton spectrum was investigated to see if any stable peaks would allow for temperature determination. Reproducibility and repeatability for three scanners from three vendors were also assessed.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;A spherical 17 cm container filled with 40% PVP w/w in distilled water was used for the experiments. A Point REsolved Spectroscopy Sequence (PRESS) with water suppression was employed on three 3T scanners from different vendors-GE, Siemens, and Philips. Frequency separation (in ppm) between peaks was measured in a voxel at the location of a fiber optic temperature probe and mapped to probe measured temperature. The center peak of the first methylene proton triplet closest to water peak was selected for analysis in jMRUI due to its ease of identification and echo time shift invariance. Shift in ppm of the central methylene peak proton was mapped against measured temperatures. Repeatability and reproducibility across the three scanners were determined at room temperature using 10 repeated PRESS scans. MRS established ppm shift versus temperature relationship was used to predict temperature in different PVP phantoms which were then compared against fiber optic probe measured temperature values.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Several &lt;sup&gt;1&lt;/sup&gt;H peaks were identified on all scans of the PVP phantom. The water peak moved by ∼-0.01 ppm/°C on the three scanners relative to a central methylene peak. The maximum mean absolute temperature difference over a temperature range of 18-35°C between the three scanners was 0.16°C while the minimum was 0.057°C. Repeatability on each scanner was excellent (std range: 0.00-0.14°C) over 10 repeated PRESS scans. Reproducibility across the three scanners was also excellent with mean temperature difference between scanners ranging between 0.1 and 0.4°C. Temperature values from MRS were within prediction bounds on the three scanners for another in-house prepared 40% PVP phantom (maximum difference&lt;0.3°C), while they were consistently overestimated for another 30% PVP phantom (&lt;1°C) and underestimated for a CaliberMRI 40% PVP phantom (&lt;2.8°C).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;PVP solutions exhibit stable proton peaks, one of which was used for assessing the temperature of the solution using MR proton spectroscopy. These measurements are fast and feasible with standard sequences and postprocessing MRS software and provide fundamental information fo","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143443128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}